1. Field of the Invention
This invention relates to a bidirectional light-activated thyristor actuated by a couple of photo-trigger means for bidirectional switching operation.
2. Description of the Prior Art
A bidirectional thyristor for effecting a switching operation in two directions in response to a signal from a photo-trigger means generally has a semiconductor substrate consisting of five layers of alternately opposite conductivities. Of these five layers, a first layer is in contact with a second layer, the second layer with a third layer, the third layer with a fourth layer, and the fourth layer with a fifth layer. The first to fourth layers with the first layer as an end layer on one side of the substrate make up a first 4-layer region, while the second of fifth layers with the fifth layer as the other end layer on the other side of the substrate constitute a second 4-layer region. Such a bidirectional thyristor has a first main electrode in ohmic contact with the outermost first layer and the second layer which is an intermediate layer adjacent to the first layer, a second main electrode in ohmic contact with the outermost fifth layer and the fourth layer which is another intermediate layer in contact with the fifth layer, and means for applying trigger signals. When such a thyristor is supplied with a trigger signal, in such a condition that a voltage is applied between the first and second main electrodes with one main electrode higher in potential than the other, one of the 4-layer regions whose forward direction coincides with the direction from the one main electrode to the other begins to conduct. When a trigger signal is applied in such a condition that a reverse voltage is applied between the main electrodes with the other main electrode higher in potential than the one main electrode, by contrast, the other of the 4-layer regions whose forward direction coincides with the direction from the other main electrode to the one main electrode enters a conductive state, thus the thyristor performs bidirectional switching operations.
Such a bidirectional thyristor is electrically equivalent to a couple of reverse-blocking thyristors (hereinafter merely called the "thyristors") connected in inverse-parallel and has the following advantages over the latter: (1) Since it may be contained in a single package, the number of cooling fins is reduced to a half, making possible compact device; (2) It is switchable in two directions in response to a trigger signal, thus offering a compact control device; and (3) No special wiring is needed for inverse-parallel connection. Because of these advantages, such bidirectional thyristors are steadily replacing thyristor couples connected in inverse-parallel in many fields of application. In spite of their advantages, practical application of such bidirectional thyristors poses problems as mentioned below.
A first problem of such bidirectional thyristors is erroneous firing in commutation. Since such a bidirectional thyristor has a semiconductor substrate which includes a couple of thyristor portions connected in inverse-parallel in proximity to each other, one of the two thyristor portions may be erroneously fired prior to the application of a trigger signal thereto by accumulated residual carriers in the other thyristor portion at the time of communication when the other thyristor portion will change from a conductive to a non-conductive state and the one thyristor portion the opposite way. Such an erroneous firing is likely to occur when the circuit to which such a bidirectional thyristor is applied has a heavy load, or when a high voltage or a high frequency is involved. Therefore, such a bidirectional thyristor is not practically useful, or even if practically useful, will find very limited application, without solving the problem of the erroneous firing in commutation.
A second problem is how to attain a substantially equal gate sensitivity of the two thyristor portions. With the two thyristor portions integrated in inverse-parallel within a semiconductor substrate, the gate electrode is inevitably disposed in one side of the semiconductor substrate. One of the thyristor portions, the PN junction of which is located farther from the side of the substrate provided with the gate electrode, has a lower gate sensitivity than the other thyristor portion. In the event that there is a wide difference in gate sensitivity between the two thyristor portions of a bidirectional thyristor, it fails to switch bidirectionally in response to the same gate signal, or even if it so switches, it is of no practical use because of the great difference in their initial conduction area. If the same switching characteristics are to be obtained in both directions in the presence of a difference in gate sensitivity, either different gate signals must be used or the gate electrodes must be differently constructed. This is practically difficult. As a result, such a bidirectional thyristor cannot be used in place of two thyristors connected in inverse-parallel unless the gate sensitivity of the two thyristor portions is caused to be substantially equal.
Apart from the aforementioned problems which are common to such bidirectional thyristors in general, another problem of erroneous firing is posed when such a bidirectional thyristor is used in an inductive load circuit. A thyristor converter is utilized in the conventional control systems for controlling the speed of a DC motor requiring accurate control operation like the motor used as a drive unit for a rolling mill. Generally, in such a control system six thyristor unit devices are provided which are connected like a three-phase full-wave rectifier circuit, each thyristor unit device being composed of a couple of thyristors connected in inverse-parallel. By replacing each of the thyristor unit devices a bidirectional thyristor, the system as a whole as well as the control circuit is reduced in size and the wiring work is saved conveniently. In spite of this advantage, the replacing of the thyristor unit devices of the thyristor converter with bidirectional thyristors poses a problem specific to the inductive load circuit in addition to the erroneous firing in commutation and the unbalanced gate sensitivity. A DC motor is always accompanied by a counter electromotive force during the operation thereof, and for activating one of the thyristor portions of such a bidirectional thyristors, it is necessary to apply a gate signal during the period when the forward voltage applied to the one thyristor portion is higher than the counter electromotive force. When the DC motor changes its rotation from forward to reverse direction or the load thereon is reduced suddenly, the counter electromotive force increases compared with that during the normal operation and, in addition, the lead angle for phase control is required to be retarded for deceleration, so that the gate signal is applied with a forward voltage lower than the counter electromotive force being applied to the thyristor portion. In other words, the gate signal is applied under such a condition that the thyristor portion to be fired is reversely biased, while the other thyristor portion connected in inverse-parallel to the former is forward biased. In this case if the bidirectional thyristor is subjected to the switching operation in two directions in response to the same gate signal or in response to a couple of synchronized gate signals from the same control circuit, the thyristor portion connected in inverse-parallel to the thyristor portion to be fired is undesirably fired, thus leading to the short-circuiting of the load and the power supply. On method to avoid this situation consists in subjecting the bidirectional thyristor to a switching operation in a range where the counter electromotive force is lower than the source voltage. This method, however, greatly restricts the speed control of the DC motor, thus making the practical application thereof impossible.
In recent years, the electric trigger system for performing the switching operation of a thyristor or a bidiretional thyristor by a supplying a gate current from the gate electrode has been gradually replaced by a photo-trigger system whereby the thyristor or the bidirectional thyristor is subjected to switching operation by photocurrent generated by radiation of light on a semiconductor substrate of the thyristor. A five-layer bidirectional thyristor of the photo-trigger or light-activated type is disclosed, for example, in the U.S. Pat. No. 33,422,323 entitled "Five-layer Light-Activated Semiconductor Device Having Bevelled Side," and granted to P. J. Whoriskey on January 14, 1969, and the U.S. Pat. No. 3,943,550 entitled "Light-Activated Semiconductor-Controlled Rectifier" and granted to Nobutake Konishi, et al., on Mar. 9, 1976. As compared with the electric trigger system, the photo-trigger system has the advantages that the easy insulation of the gate circuit from the main circuit simplifies the gate circuit construction and at the same time facilitates application to a high-voltage circuit, and that the absence of feedback from the main circuit to the gate circuit eliminates the erroneous firing which otherwise might occur due to the inductive action. The photo-trigger system, in spite of these advantages, has various problems yet to be solved as mentioned above, if it is to be applied to the bidirectional thyristor.